The Alternative Vote - the system no-one wants by Policy Exchange

research note October 2010

September 2010

The Alternative Vote – the system no‐one wants Dr Robert McIlveen. Edited by Natalie Evans Dr Robert McIlveen

Executive Summary In 2011 the British people will have their say on whether to change the electoral system used to elect MPs to the House of Commons to the Alternative Vote (AV). The Deputy Prime Minister has described this choice as part of a package to reduce the “deep unfairness” in how our system works at present, where one party can win a large majority on a small lead but others cannot win a majority on a much larger lead – and third parties struggle to win seats anywhere near their share of the vote. AV is only used for general elections in Australia, Fiji, Papua New Guinea and Nauru, although all of these are different versions to that proposed for the UK. Fiji has recently decided to abandon AV in favour of Proportional Representation (PR) In Australia it was introduced for partisan interest: to prevent Labor winning seats where the right‐of‐centre vote was split between the various conservative‐inclined parties. There are many claims about AV and how it might improve on our current electoral system, First Past the Post (FPTP). This report examines the case for change and whether AV does genuinely improve on FPTP or if it is reform for the sake of it. The most common charge against FPTP is that seats in parliament do not reflect national shares of the vote. This is true but is so for any system that prioritises having a single MP for each constituency, including AV. Academic evidence suggests that AV can be much more disproportional than FPTP. It might even increase the bias in the system to Labour, if the other sources of bias (turnout, efficiency and constituency size) continue: •

It would have produced even larger majorities for Labour in 1997 (213 or 245 compared to 179, according to two studies) and in 2005 – when the Labour Party achieved only 35.3% of the UK vote, it would have had a majority of 108, compared to 66 under FPTP.

•

In 2010 – when Labour scored only 29% of the vote – it would have delivered them almost as many seats (248) as the Conservatives, on 36.1% of the vote (283).

•

In terms of perverse results – when the party winning the most votes does not win the most seats – AV’s track record in Australia is terrible, with three elections between 1969 and 1998 delivering a perverse result. In the same period it happened once in the UK, in February 1974. In 2010 Labor secured fewer first preference votes than the Liberal‐National Coalition but still managed to put together a coalition government.

Supporters of electoral reform claim that FPTP “wastes” votes because many people’s vote does not elect the winner. This is an argument designed purely to support the Single Transferable Vote (STV), the “all must have prizes” electoral system, and does not work well for AV: •

Under AV, plenty of votes would still be “wasted” – in a seat won on 50.1% of the vote, 49.9% would still be “wasted”

•

AV reduces wastage by recycling votes, however this makes one voter’s second or third (or even fourth) preference equal to another voter’s first preference.

AV ensures that MPs have at least 50% of the vote after recycling, but this is not necessarily full‐blooded support. It solves a problem no‐on cared about before the choice was narrowed to FPTP or AV – and indeed is a test failed by the system most pro‐reform campaigners prefer – STV. The Jenkins Commission made an important distinction between support for a candidate and acquiescence to their election, which AV conflates into full‐blooded support. It is questionable whether an MP elected on 50.1% after third preferences is more legitimate than one elected on 49% under FPTP. Tactical voting is another element of the reform case, with critics claiming that people are forced to vote tactically under FPTP and that this is wrong. AV formalises tactical voting so everyone does it but at no cost in terms of the ultimate outcome: you can have your cake and eat it too, potentially several times. The case against tactical voting underestimates the sophistication of the electorate and prejudges their motives for voting. Tackling seats for life is one of the most important items on the reform agenda. MPs’ expenses over‐claims were statistically linked to length of service but not to the safeness of an MP’s constituency as is often supposed. There are much better ways of removing under‐performing or dodgy MPs: recall ballots are also being proposed which is a big step in the right direction. The political parties could also be much more effective at rooting out weak or dubious MPs. In any case, AV does precisely nothing about safe seats, which would continue to be ignored at election time and potentially taken for granted by parties and MPs. One interesting difference between the UK and Australia is the difficulty minor parties have in breaking through: •

The Australian Greens won their first seat in the House of Representatives in 2010, but gained only one seat (out of 150) for a vote share of over 11%.

•

In the UK under AV, the Greens and the Scottish and Welsh Nationalists might struggle to hold on to their seats, none of which were won with more than 50% of the vote.

AV is the system no‐one actually wants. Supporters of reform usually prefer STV or the Mixed Member Proportional (MMP) system in use in Scotland, while constitutional conservatives see significant danger in a switch to AV. The appeal of AV is limited and is largely down to those who think it might be a first step to a more radical reform, which risks permanently saddling us with a system that is worse than almost all others available.

Recommendations The case for changing the electoral system to AV is extremely weak and should be rejected. It only improves unquestionably on FPTP in one, trivial, regard – MPs must have 50% acquiescence from their electorate – but at the considerable cost of treating voters unequally by recycling votes. The electoral system for the House of Commons should be decided on the merits of each system under discussion, not on the basis of future party realignments and further reform. While AV as a slippery slope may be a strong motivation for committed reformers it risks saddling the UK with an electoral system no‐one actually wants. If retained, FPTP should be reformed to be as unbiased as possible. While most of the pro‐Labour bias in the electoral system is due to party and voter behaviour, unequal seat sizes must be addressed since they do contribute to some of the bias. Policy Exchange will consider this in more detail in future research. If the referendum on AV does pass, the change should be permanent. Any further reform would dilute or destroy the constituency link and should be ruled out.

Introduction: Why are we having a referendum? On 5th July 2010, the Deputy Prime Minister outlined the Government’s political reform programme. Proposals to reform the electoral system, reduce and equalise constituencies and other changes would, he argued: “help to correct the deep unfairness in the way we hold elections in this country. Under the current set‐up, votes count more in some parts of the country than others, and millions feel that their votes do not count at all. Elections are won and lost in a small minority of seats. We have a fractured democracy, where some people's votes count and other people's votes do not count; where some people are listened to and others are ignored.”1 The “deep unfairness” by which some votes are more important than others is a part of how elections have operated for a long time in this country. In 1983, the Liberal‐SDP alliance won 25.4% of the UK vote but only 23 seats out of 633, which left a lasting impression on the third party. The practice of focusing on swing voters in swing seats is not new. The claim that we have a “fractured democracy” may well be parliamentary hyperbole; but it also taps into the dissatisfaction apparent throughout the expenses scandal. The uncertain outcome of the 2010 general election on top of the low standing of Westminster’s elites with the public raised doubts about our political system’s durability. Politicians will normally try to calculate if there is a partisan advantage to electoral reform. AV might benefit one party over the others, but we can fairly assume that the parties would realign after reform anyway,

making such predictions extremely hazardous. In any case, electoral reform should not be conducted on the basis of partisan interest but with the aim of improving our democracy – the approach this Research Note takes. In an attempt to address the perceived problems with our system, the government is considering how we draw constituency boundaries and how we vote. The two key elements are the Alternative Vote (AV) referendum and boundary changes to create equal‐sized constituencies which are connected in the government programme. A referendum on changing the voting system is a product of compromise in the creation of the coalition. The Liberal Democrat manifesto pledged support for the Single Transferable Vote, a much more radical reform than that being offered, while the Conservative manifesto was steadfastly opposed to any reform of the electoral system. AV is a strange choice for an alternative system, since it offers only minimal reform. It is unlikely to be adequate to “correct the deep unfairness” in the electoral system. One key argument for reform is the bias towards the Labour Party in the electoral system as it stands. In 2005 a Labour lead of 2.8% produced a 66 seat majority, while in 2010 a Conservative lead of 7.1% fell 19 seats short of a majority. Understanding why the electoral system is biased to Labour is important, not least because it impacts on nearly every aspect of the government’s reform programme. Electoral reform was last on the agenda around the 1997 General Election, in which the Labour manifesto promised a referendum. The Jenkins Commission which was established following that election duly recommended AV+, a hybrid system not in use anywhere. After the election momentum for reform dissipated somewhat; and despite commitments to a referendum in subsequent Labour manifestoes it never happened and the Jenkins Commission’s report was left languishing. The views of the Jenkins Commission on AV are well worth noting. The report rejected AV on the grounds that: “First, it does not address one of our most important terms of reference. So far from doing much to relieve disproportionality, it is capable of substantially adding to it. Second, its effects (on its own without any corrective mechanism) are disturbingly unpredictable. Third, it would in the circumstances of the last election, which even if untypical is necessarily the one most vivid in the recollection of the public, and very likely in the circumstances of the next one too, be unacceptably unfair to the Conservatives.”2 Demand for electoral reform has emerged at various times in the UK’s recent political history, although almost always as an elite interest rather than mass surges of popular discontent. MORI’s long‐running series asking the public to name important issues facing Britain demonstrates almost total indifference to the admittedly very broad category of “Scottish/Welsh Assembly/Devolution/Constitutional Reform”. The proportion of the public mentioning this as an important issue did not exceed 2% between May 1997 and June 2010 and was often not mentioned at all.

Principles for reform No electoral system is perfect, and all choices, including First Past The Post, (FPTP) require decision makers to trade off between different values. Many of the issues around electoral reform were considered by the Jenkins Commission in 1998, which was tasked with recommending a new electoral system for the UK and was required to “observe the requirement for broad proportionality, the need for stable Government, an extension of voter choice and the maintenance of a link between MPs and geographical constituencies.”3 Other key requirements for an electoral system implied by Mr Clegg include the equal weighting of votes and voters. The key question for the referendum is whether AV is an improvement on FPTP. This argument will revolve around several key issues. This note will compare AV and FPTP on six key tests: •

proportionality;

•

safe seats;

•

decisive results;

•

wasted votes;

•

tactical voting; and

•

MPs being elected on less than 50% of the vote.

These six cover issues key to the two systems as well as those which have been given prominence in the debate around how we elect our MPs. How to choose between the two depends in large part on how much weight is placed on each factor, and whether they deserve to be taken seriously in the first place.

What is the Alternative Vote? A different version of the Alternative Vote (AV) system proposed for the UK is used in Australia to elect MPs to the House of Representatives (the difference is that voters must rank all candidates, whereas in the UK version they may rank as many or as few as they wish). A weighted version of it is used in Nauru, while Papua New Guinea uses a limited version of it and Fiji uses it but with reserved seats for different ethnic groups, making comparison difficult. There are some claims that AV is particularly suited to ethnically divided societies such as Fiji and Papua New Guinea since it tends to produce moderate outcomes, although these claims are contested.4 That Fiji is planning to abandon AV before their next election does not bolster the case for AV. The system works by having voters rank candidates rather than simply supporting one, as in FPTP. It retains single‐member constituencies. To win, a candidate needs the support of 50%+1 of all votes cast. Where this is achieved on first preferences, the candidate is elected straight away, so in this regard is functionally no different to FPTP. Where the first count does not produce a winner with 50%+1, losing candidates are eliminated and their first preferences redistributed to each of their supporter’s second preference. This carries

on until one candidate has achieved 50%+1. Second, third and lower preferences are counted equally with first preferences once they are redistributed.

How it works in Australia AV has been used in Australia for elections to the House of Representatives since 1919, where it is known as preferential voting. It was introduced by the Nationalist Party following the Swan by‐election in 1918, when Labor won a usually safe Nationalist seat because the conservative‐inclined voters split between the Country and Nationalist candidates.5 Two versions of it are in use today: ‘full preferencing’ where voters must rank all candidates to enter a valid vote which is used for general elections to the House of Representatives; and ‘open preferencing’ where they rank as many as they wish, which is used in some state elections. This is politically important because the Labor party is seen to be disadvantaged by full preferential voting, as it allows conservative‐inclined votes to be redistributed amongst several conservative parties such that one of them wins even in situations where Labor leads strongly in the first round. An extreme example of this is the Riverina constituency in 1980, where Labor lost despite having the support of 46.9% of the electorate on first preferences. Table 1. Riverina Constituency, 1980 General Election6 Party

1st Round

2nd Round

3rd Round

4th Round

Result

Labor

46.9%

47.1%

48.3%

49.5%

Defeated

National Country 33.1% Party

33.5%

34.4%

50.5%

Elected

Liberal

16.9%

17.0%

17.3%

Excluded

Australian Democrats

2.2%

2.4%

Excluded

Independent

0.8%

Excluded

AV means that the fragmentation of parties on the right, now between the Liberals and Nationals, does not split the vote allowing Labor to win. This paper will treat the Liberal‐National coalition in Australia as a single party since the two parties very rarely compete against each other in elections to the House of Representatives, except in safe seats when the sitting member retires.7 As in the UK, the 2010 Australian general election was unusual in that it resulted in the country’s first hung parliament since 1940. The two major blocs retained their very high share of the seats – only 5 out of 150 were not won by Labor or the Coalition. Interestingly the results demonstrate the disproportionality of elections under AV, with the Coalition’s 5.6% lead (on first preferences, at time of writing) not giving them the most seats. The Greens, who won 11.76% of the vote, only gained one seat.

Minor parties had been notable by their absence in the Australian House of Representatives until the Greens’ victory in Melbourne in 2010. The Senate, elected under STV, has seen minor parties succeed, including the Greens, Family First Party and Australian Democrats and One Nation. While independent MPs are slightly more common in the house of Representatives than in the UK these are typically disgruntled former major party candidates, who in the UK have not typically held their seats. A worked example in the UK: how Reading East would have turned out under AV in 2010 Reading East was gained by the Conservatives in 2005, with Rob Wilson becoming the first MP to have been selected by an open primary. In 2010, the Conservatives held the seat, gaining 42.6% of the vote with a majority of 15.3%. In Sanders et al, Reading East is one the more surprising seats which they simulate would have changed hands under AV. This is due to the way the votes would be redistributed, based on their survey evidence. The key transfer is Labour to Lib Dem in the final round, because 70% of Labour second preferences would go to the Lib Dems and only 6% to the Conservatives. This is based on the British Election Study data, which asked respondents to fill in a simulated ballot paper for an election run under AV. How Reading East would have worked out under AV % of the vote for each party over 4 rounds

Round 1

Round 2

Round 3

Round 4

Con

42.6%

43.0%

44.5%

49.4% Defeated

Lib

27.3%

28.5%

28.9%

50.6% Elected

Lab

25.5%

26.1%

26.6%

UKIP

2.2%

2.4%

Green

2.1%

Eliminated

Ind

0.2%

Eliminated

Ind

0.1%

Eliminated

NB: due to lack of data on preferences, Independents’ votes disregarded after first round Interestingly, since Lib Dem voters split evenly between the Conservatives and Labour, at least in the South East of England, the Conservatives would win in a scenario where Labour are their opposition in the final round.

FPTP vs. AV On some factors, AV and FPTP are indistinguishable – both have single‐member constituencies and a change would make no difference to the constituency link. On otjhers, AV is a dilution of FPTP – for example, it would make dismissing the government harder in the UK unless the parties realigned back into a two‐party system. The main issues for comparison of FPTP and AV are: disproportional outcomes (in terms of share of vote and share of seat each party receives), the existence of safe seats, decisive results, the “wasted votes” argument

and tactical voting. In addition to these long‐standing issues, since the announcement of a referendum on changing from FPTP to the Alternative Vote (AV), much more attention has been given to a sixth argument: MPs elected on a share of the vote less than 50%. This section explores these six factors, considering whether they actually matter, and whether a change to AV would improve matters substantially. 1. Bias, disproportionality and perverse results The biggest criticism of FPTP is that the results of elections held under it tend not to distribute seats in the House of Commons proportionally to national totals of votes received by each party. The 2005 election has been seen as particularly unfair – the Labour party won only 35.2% of the vote but held 55.2% of the seats in the House, to the Conservatives’ 32.4% votes and 30.7% seats. Figure 1. Seat:vote ratios, 1970‐2010

1.8

% of seats for every % of vote share

1.6 1.4 1.2 1.0

Con seat: vote ra�o

0.8

Lab seat: vote ra�o Lib seat: vote ra�o

0.6 0.4 0.2

10 20

05 20

01 20

97 19

92 19

87 19

74 19

(O )

) (F 74 19

Figure 1 shows the historical trend for the three parties as a ratio of their percentage share of seats to their percentage share of votes (the seats:votes ratio). Prior to 1997, it was normally the case that the winning party had a better ratio than the losing party, but that neither of the big two parties was substantially disadvantaged compared to the other. In 1997 this changed significantly, with the Conservatives losing far more seats than their share of the vote would have suggested alone. Strikingly, in 2010, the Conservatives still had a smaller seats:votes ratio than Labour – 1.30 to 1.37 – despite being the “winning” party. This is the only time since 1951 that the party with the most votes has not been the party with the best seats:votes ratio.8

Disproportionality is due to a mix of factors, including constituency size, different levels of turnout in different types of seats, and some parties’ support being more efficiently spread than others. Some of these are down to voters’ behaviour and cannot be tackled through electoral reform, such as turnout being lower in more deprived seats than in more prosperous seats. Similarly, part of Labour’s exceptionally good performances since 1997 are down to its efficient spread of votes – it has been very good at concentrating its efforts where it needs to. Their ruthless focus on target seats and implicit anti‐Tory co‐operation with the Liberal Democrats combined to produce an exceptionally efficient distribution of votes – something the Conservatives have struggled to match since.9 Do disproportionality and perverse results matter? Disproportionality is the most important motivation for many campaigning for electoral reform. The strongest appeal for change rests on the assumption that proportionality is something the electoral system should aim for. For example, the Jenkins Commission had “broad proportionality” as one of its four requirements. Disproportionality is the direct result of single‐member seats. It is impossible to guarantee proportionality without either sacrificing the notion of single‐member seats or introducing “top‐up” MPs as in the Mixed Member Proportional (MMP), the system used in Scotland and Wales. The latter introduces the problem of having two classes of MPs which all parties at Westminster have been keen to avoid. MMP might also have deprived the nation of the “Portillo moment” had the Cabinet in 1997 been on the safe top‐up lists rather than defending constituency seats.10 “Perverse results”, where the party with the most votes does not win the most seats, occur every so often in any single‐member system. In contrast to its supporters’ claims about proportionality, Australian elections under AV has produced such an outcome more frequently than the UK. Since 1969, there have been three occasions (1969, 1990, 1998) on which either Labor or the Coalition had more votes but fewer seats than their rivals on final preferences, and five occasions (the previous three plus 1987 and 2010) on which this is true of first preferences. Such a situation has only happened once in the UK in the same period – February 1974. The key question is does proportionality matter more than the constituency link? For many years the answer in the UK has been no. The anonymity of many MEPs in their very large constituencies is hardly an inspiring example of effective representation in multi‐member constituencies. Would AV make any difference? AV has been branded as an attempt to bridge the gap – it retains the constituency link but, its proponents claim, gets closer to a proportional outcome. However, Dunleavy et al (1998) found that had it been in use in the 1997 General Election, AV would have been significantly less proportional than FPTP. This analysis was based on a large‐scale experimental ballot following the actual election.11 On a 31.4% share of the vote the Conservatives would have received only 19% of the seats in 1997 under AV – as the authors note, “the most severe under‐representation of the Tories in British history”. Labour would have won a majority of 213

(compared to 179 in reality) according to the paper, having 68% of seats on 44% of the vote. The Jenkins Commission cites an even more extreme estimate, producing a majority of 245. In 2005, AV would have produced another extremely disproportional result, giving Labour a majority of 108 on only 35.3% of the vote, when under FPTP they achieved a majority of 66. 12 2010 would be less extreme but would still be highly generous to Labour, giving them 248 seats (or 38% of seats in the House of Commons) for only 29% of the vote.13 Essential reading on simulated results and different systems Dunleavy, Margetts, O’Duffy and Weir (1998) “Remodelling the 1997 general election: How Britain would have voted under alternative electoral systems” uses a very large survey with mock ballot papers to model how the 1997 election would have worked out under many different systems; only AV and the Supplementary Vote come out worse than FPTP. Sanders, Clarke, Stewart and Whiteley (forthcoming in 2011) “Simulating the effects of the alternative vote in the 2010 UK general election” uses a similar method to simulate the most recent election result under AV. It projects that both the Conservatives and Labour would lose out (the Tories more so) and Lib Dems gain enough to make a Labour‐Liberal coalition possible. Hix, Johnston and McLean (2010) “Choosing an Electoral System” explains a wide variety of different systems, including projected results from various sources. The Jenkins Commission (1998) is a thorough exploration of what fairness and representation mean, as well as a consideration of a wide range of systems Under AV, turnout would still be a major source of bias, since Labour safe seats tend to have very low turnout and in many cases would not be affected by a change to AV since those elected get over 50% of the vote. For example, no seat in Liverpool would have required a second round in 2010 (or indeed any election since 1992) had it been held under AV, since Labour have achieved at least 50% of the vote in all of these seats since then, on relatively low turnouts.14 Figure 2 displays the relationship between turnout, majority and incumbent party, with Labour‐held seats dominating the left‐hand side (lower turnout) and Conservative and Liberal Democrat‐held seats the right hand side (higher turnout). The trend is most concentrated in the extreme seats: only one in four of the bottom 50 seats by turnout are not Labour‐held while only 2 of the top 50 are held by Labour.

Figure 2. Party distribution of turnout and winning margin in 2010

7575 7070 6565

Winner’s % of vote Winner’s % of vote

6060 Conserva�ve Conserva�ve

5555

Labour Labour

5050

LibLib Dem Dem

4545 4040 3535 3030 2525 4040

4545

5050

5555

6060

6565

7070

7575

8080

% Turnout % Turnout

Where AV might make a difference is in how parties compete with each other. With a change in the rules, how they win seats would be different – for example, they would have to explicitly appeal for other parties’ supporters to give them their second preferences. This may or may not make results more proportional – in 1997 it would have increased the other parties’ efficiency against the Conservatives and produced an extremely disproportional result; in most elections it would boost a centrist or non‐threatening third party’s share of seats at the expense of one or both the larger two. If this is done evenly between the big two it would tend to reduce partisan bias, if it is uneven it would increase it. Conclusions on disproportionality and perverse results Disproportionality is the most important charge against FPTP, since it is unarguable that it does not elect MPs in proportion to national support for the parties. It is often asserted that AV would be more proportional than FPTP while retaining the constituency link; however the evidence suggests this is not the case. Some elections would be much more disproportionate than under FPTP, some would be less so. There may well be a partisan basis to this, with a choice between large Labour majorities and Conservative‐Liberal coalitions, although it does depend on how each party’s supporters’ votes split at each election. Disproportionality is not a valid argument for changing our electoral system to AV. Perverse results is even weaker, since Australia has these outcomes more often than the UK does, and nothing about AV would make them less likely. 2. Safe seats Take Back Parliament, an umbrella group for electoral reform, argues that FPTP “empowers a few thousand voters in "marginal" seats who decide elections, while those in "safe" seats, where the MP has a large majority, are ignored.”15 This is true: but parties will always focus on swing voters under any electoral system, and all single‐member systems make the parties focus on this group of voters in swing seats.

Defined by majority, there were many safe seats in 2010. 274 seats saw a majority of 20% or more – well in excess of all but one swing in that election (Redcar, which saw of swing of 21.8% from Labour to the Liberal Democrats). There were 212 seats where the winner received over 50% of the vote. There were 196 seats where the majority was over 20% and the winner’s share of the vote was over 50%. In total, 291 out of 650 seats have winners with either more than 50% or a majority of more than 20% or both, making them very safe. Do safe seats matter? There are several reasons safe seats are criticised. A link between safe seats and the recent expenses scandal has often been asserted, with critics arguing that MPs in safe seats faced no realistic opposition and so felt free to line their pockets. Regression analysis suggests that length of service was actually the key factor, since MPs elected in safe seats in 2001 or 2005 were no more likely than MPs elected in marginal seats in the same elections to claim dubious expenses. The regression analysis below show that there is no statistically significant relationship between the size of an MP’s majority and the amount of expenses mis‐claims, but there is a highly significant relationship between length of service and amount mis‐claimed. While this isn’t a perfect model – there’s no honesty index available for MPs and the adjusted r‐squared value, which estimates how good a description of the data the model is, is very low – it is clear that the safety of a seat is not the cause of unjustified expenses claims. Table 2. Regression of length of service and majority in 2005 on MPs’ expense repayments Variable

Coefficient (p value)

Length of Service

165.07 (0.000)***

Majority in 2005 (%)

‐24.24 (0.313)

Adjusted R‐squared = 0.05 *** p<0.001 Beyond recent scandals though, the argument against safe seats has some merit – MPs with no prospect of a challenge from other parties face little pressure and can treat it as a seat for life. Baldrick’s slogan in the Dunny‐on‐the‐Wold by‐election in Blackadder, “A rotten candidate for a rotten borough” contains a grain of truth, since MPs in safe seats can potentially afford to be much less diligent than those in marginals. They can, however, also be more independent of their party whips as long as they keep their local party on side. There are other options to make safe seats less secure. Government proposals for recall ballots in the coalition agreement would make MPs found to have committed serious wrongdoing open to a recall ballot in their constituency. While this is a small step it could have significant ramifications. Would AV make a difference? Not really. The 291 seats where the winning party received either 50% or more of the vote or a majority of at least 20% (or both) would in all likelihood remain safe seats. Unless AV radically changed voting habits, most of these seats would be won on first preferences or won easily on second preferences.

Some seats might become more winnable – for example, where one party holds the seat against fractured opposition. But these are seats which are probably not “safe” under FPTP, but are “long shots”. The impact of AV is likely to be felt in marginal seats, where second and third preferences could really come into play. Sanders et al simulated the 2010 result under AV, producing a list of 43 seats which would have changed hands from the real 2010 result, with the Lib Dems gaining 19 seats from Labour and 13 from the Conservatives; Labour gaining 10 from the Conservatives and the Conservatives gaining one from Labour. These are almost all marginal seats which would be competitive under FPTP anyway, such as Chesterfield, Cardiff North, Colne Valley and Dudley North. The few exceptions were seats where the Liberal Democrats were in second with either the Conservatives or Labour in third with a large share of the vote (e.g. Lewisham West and Penge or Reading East).16 It is even possible that different types of safe seats could emerge, for example if two parties locally form a stable agreement to recommend each other as second preferences – not unlike the history of electoral pacts between the Conservatives and Liberals in the 1950s, under which four Liberal seats were not contested by the Conservatives to help keep Labour out.17 Under AV parties could play the system through pacts or even formal coalitions as in Australia. This would increase the number of safe seats, as long as most supporters of the parties involved continued to support a coalition. Conclusions on safe seats Criticisms of safe seats are not new and are common to all single‐member systems, including AV. At the last election, the Conservatives introduced a variety of different ways to select candidates in safe seats, while Labour have used All‐Women Shortlists. Open primaries, especially the all‐postal version used in Totnes and Gosport are particularly impressive, albeit very expensive. While these do not solve the problem of a seat for life, they do suggest that parties can be creative about selecting MPs, and that we do not need to change how votes are counted to change how secure MPs are. The Coalition Government’s agreed programme includes a commitment to enabling voters to remove their MP if he or she is “found to have engaged in serious wrongdoing”, effectively a recall procedure (as used to remove the Governor of California, Gray Davis, in 2003) albeit with a high threshold.18 Calls to strengthen this have already been made and could result in a radical assault on seats for life.19 AV does very little to tackle safe seats, since it has much more impact in marginal seats. Australia has safe seats under AV – Gellibrand, Victoria gave Labor 74.21% of final‐round preferences, while Mallee, Victoria gave 74.4% to the National Party in 2010. In 2010, 59 out of 143 seats that were straight Labor/Coalition fights had majorities after second preferences of over 60%, 9 of over 70% making them extremely safe. This is similar to results in the UK in constituencies such as Liverpool Walton (72%) or East Ham (70.4%). Under almost any electoral system, safe seats are possible – even under STV in some circumstances, such as Northern Ireland’s elections to the European Parliament where each party has one candidate and the two main unionist parties have won a seat each in every single election.

3. Decisive results The strongest case for FPTP is that it makes it easy to “throw the rascals out” – something a coalition‐inducing system makes harder. Post‐war West Germany is an excellent example of how difficult it can be to get rid of a party in government: the FDP was in government for all but three years between 1961 and 1998. This means that FPTP is the best guarantor of accountability. In the West German case, the Free Democratic Party really faced no proper accountability to anyone. Indeed, their only departure from government was caused by the resignation of FDP cabinet members over the 1966 budget and the Christian Democratic Union turning to a grand coalition with the Social Democratic Party – so effectively by their own choice. FPTP is often supported by the argument that it fosters strong, decisive government, since it rarely delivers coalitions. While it does occasionally produce minority or coalition governments it usually produces a majority. It also allows for a second election (as in 1951, 1966 or October 1974) when an election does produce a tight result, which usually results in a more decisive outcome. This is generally contrasted to weak coalition governments in countries such as Israel, Italy (historically) or the Netherlands which have more proportional systems and more frequently collapse, often under demands from junior partners. Do decisive results matter? Yes – the Westminster system is predicated on governments with clear mandates. While the UK has had coalition governments, and has one at the moment, these are relatively rare outside times of war or crisis. Would AV make a difference? Yes, if we assume that voters and parties do not change their behaviour. Under AV, had voters voted exactly as they did in the actual elections, we would have seen coalitions, minority governments or fresh elections in 2010 and probably in 1992 as well. But this is not a good assumption at all, since parties are very good at strategic realignments and voters are sophisticated. Both have already learnt to factor in second preferences (London Mayoralty) and regional top‐ ups (Wales, London Assembly and Scotland) and to operate in closed‐list elections (European Parliament). The Australian example gives an idea of how UK elections under AV could work out. There are essentially two plausible governments – Labor or the Liberal‐National Coalition. Australia has far fewer small parties represented in Parliament than the UK does – the Greens are the first small party to break through, although there are more independent MPs than in the UK. And they needed a large share of the vote to gain just one seat – compare their 11.7% and one constituency to the UK Greens, who gained one seat with only 1% of the vote. Would something like this happen in the UK? It is quite possible that party realignment could occur, but would it be more or less fragmented? One possibility is that the Welsh and Scottish nationalists and the Greens could

be squeezed out by the big three parties, since in none of the ten seats won by the minor parties in 2010 did they gain more than 50% and in seven they received less than 40%. Conclusions on decisive results In terms of decisive results, AV is rather odd. It would probably have produced even larger Labour landslides between 1997 and 2005, but made it harder for the Conservatives to win a majority on a large lead. This is not the direct result of AV but the biases inherent in the electoral system, which AV would likely exaggerate. The decisive results argument for FPTP relies on the system being decisive but not biased. Changing to a system which exaggerates swing without trying to reduce bias in the system risks making the system more biased. 4. Wasted votes Emotively, this is the most appealing of electoral reformers’ arguments: that “FPTP in effect wastes huge numbers of votes, as votes cast in a constituency for losing candidates, or for the winning candidate above the level they need to win that seat, count for nothing.” 20 This is usually used as an argument in favour of STV, a more complex system which reallocates votes in multi‐member seats over several rounds. Whether used in aid of AV or STV, the wasted votes argument relies on the idea that a vote for a losing candidate does not count, and that recycling it to that voter’s second (or third) preference makes it count. The argument for AV and STV also makes the very strong assumption that a second preference is the equivalent of a first preference. This assumes that a vote is only important as a means to elect a winner in an election, which is a very strong assumption – motivations to vote are very complicated, especially since it is almost never rational to bother voting at all if all that matters is having a decisive effect on picking the winner.21 Voters who strongly support a party in the knowledge that it will lose may not want to support another as a second preference but simply want to register their support; some will vote on single issues; some randomly; some habitually. Do Wasted votes matter? The wasted votes argument is predicated on the idea that as many voters as possible should be represented by someone they support, or else their vote is wasted. An extension to this runs that even votes for winning parties are wasted as “surplus” votes since the winning party does not need them. The Electoral Reform Society supports changing to STV because under it “most voters can identify a representative that they personally helped to elect. Such a link in turn increases a representative's accountability.”22 While this overlooks the large number of non‐voters, the link between electing someone and them being accountable is dubious too – I could hold my MP accountable despite having not voted for her, either through the ballot box at the next election, the law or even turning up at her constituency surgery.

The wasted votes argument assumes that MPs represent their constituents personally and are doing this illegitimately for their constituents who did not vote for them, which seems rather at odds with the traditional conception of the MP as representative of his or her constituency as a geographic community– not just those voters who elected him or her. One academic consulted pointed out that “All votes are wasted in the sense that no single vote affects who is elected”. Wasted votes is an argument designed to support STV, which is effectively the “all must have prizes” approach to elections. This only works if you make the strong assumption that the only thing that matters when voting is to have a decisive impact on selecting the winner, which is virtually impossible in large electorates. If you value any other reasons for voting, this argument fails. Would AV make a difference? Under AV significant numbers of votes would still be wasted, since it is a single‐member system. Any seat where the winner gains 50.1% would see 49.9% wasted. Some which are currently wasted would be recycled, while others which currently count would become wasted as a result of such recycling, in seats where the outcome would be changed by reallocating votes. AV’s (partial) solution to the wasted vote could fairly be described as a recycled vote – one where a voter’s second, third or even fourth preferences are counted equal to other voters’ first preferences. This is the biggest flaw in AV (and STV): the equal weight given to first and subsequent preferences. Why should a twice‐ recycled vote (say Green to Liberal Democrat to Labour; or English Democrat to UKIP to Conservative) count as equal to a first preference for another party? Conclusions on Wasted votes Votes for a losing candidate under any system do count – they just don’t win. Wasted votes is a superficially attractive argument, but is unconvincing. Elections produce winners and losers, and the supporters of the losers are not disenfranchised or ignored – they just supported the losing candidate. MPs represent their entire constituency, not just those who voted for them, or just those who voted at all. Recycling votes through AV is more unfair than people voting for losing candidates under FPTP – which would still occur under AV in any case. Each person’s vote should count equally, which is not true under AV since some voters’ third preferences are treated equally to others’ first preferences. 5. Tactical Voting Voters who want to are sophisticated enough to respond to the “wasted votes” idea in any case – by tactical voting. Tactical voting under FPTP is when a voter votes for their most preferred candidate out of those who they think can realistically win the seat, rather than their most preferred candidate who they think will not. Those who consider it face a choice between conviction and pragmatism under FPTP. Supporters of AV argue

that people should not have to do this, and that AV means that they can sincerely vote for their first preference before ranking the other candidates, meaning an end to tactical voting. The above argument is almost entirely correct – under AV voters can vote for someone they like the most before voting for their preferred candidates out of those who are in contention to win. In this way, AV removes the need to not vote for your favourite candidate. However, it replaces such situations with everyone (at least those who are not so strongly partisan that they only vote for one candidate) voting tactically through their second and third preferences. So it changes the nature of tactical voting by formalising it within the electoral system. There are still occasions under AV in which voters have an incentive to vote for someone other than their most‐preferred candidate – what we understand as tactical voting under FPTP. Where the result between second and third is likely to be close, the potential for AV tactical voting emerges. Tactical voting under AV can occur if voters are trying to determine who the final round is between – for example, Labour would be more confident of beating a Conservative in the last round than a Liberal Democrat in most circumstances, so might be keen to have a small number of supporters vote Conservative in the hope of tipping them over the line into second place in the penultimate round before beating them on Liberal Democrat voters’ second preferences. As an example, in the 2001 Conservative Party leadership election, conducted under multiple knock‐out rounds similar to AV, it is alleged there was tactical voting in the final MPs‐only round to ensure that Iain Duncan Smith faced Ken Clarke who he was confident he would beat, not Michael Portillo who he was less confident about. It would be possible for this to occur under AV in a general election, although it would require quite a significant amount of organisation to pull it off. Does tactical voting matter? The argument about tactical voting is similar to the wasted votes discussion about voters’ motives. The assumption made by the reformers is that voters are forced to vote tactically by the system rather than choosing to do so because they are sophisticated and particularly want to stop another candidate. Tactical voting is a fact of life under FPTP, and one that has been grasped by all parties in their campaigning. But it is not clear that it is bad or immoral – voters should be free to vote how they like for whatever motive they choose. The argument that tactical voting is bad and AV removes the need to do it presupposes the first element, which in turn suggests a rather judgemental and patronising attitude to voters. Would AV make a difference? The move to AV would effectively allow people to vote for their least‐disliked party without having to not vote for the party they actually like – or having their cake and eating it, perhaps twice (our Green‐Liberal Democrat‐ Labour voter, for example). This formalises tactical voting, allowing everyone to do it at no actual cost in terms of the result.

AV would be more permissive of minor parties’ support being recognised in the first round, although probably less likely to actually elect any minor party MPs. It would empower staunchly eurosceptic Conservatives, for example, to vote UKIP safe in the knowledge that they can transfer back later. In this sense it allows voters more room for expression, especially with single‐issue parties, which we might expect to flourish in terms of first preferences. This is where arguments about electoral reform and extremist parties come in. AV offers an opportunity for voters who want to register a protest to vote for an extremist or fringe party and then transfer back to a mainstream party having expressed their protest. We might expect the BNP (as well as UKIP and Greens) to gain significantly more first preferences under AV than they do now when the cost of voting for a party that will not win is significantly higher. Expressive voting currently carries a high cost, but would become effectively free under AV, making it rather more frivolous. Conclusions on tactical voting It is unarguable that AV largely removes the need for tactical voting as we know it under FPTP. If you believe tactical voting to be a bad thing this makes the case for AV stronger; if you take a less patronising view of the electorate it simply changes one way to make a decision into another. AV just formalises tactical voting to the point where everyone does it. 6. MPs being elected on under 50% of the vote The only way in which AV is definitely superior to FPTP – and indeed superior to more proportional systems such as STV – is that all MPs would be elected with the support of over 50% of those who voted. FPTP does not attempt to guarantee this – in Norwich South in the 2010 election, the winner had a share of only 29.4% in a tight four‐way marginal. Very low winning vote shares tend to be in the most competitive seats, especially where three or four parties are in contention. As Figure 3 shows, the last two elections have seen an unusually high proportion of MPs elected with less than 50% of the vote. In both cases around two thirds of seats saw a winner with less than 50%, compared with fewer than half in 1997. This measure is basically the same as asking how much support the third party in each seat got, since their success brings down the share of the top two parties – and hence increases the likelihood of the seat being won with less than 50% support. Since the Liberal Democrats are most often the third party, their vote share is unsurprisingly very closely correlated with this phenomenon.

Figure 3. The proportion of MPs elected with less than 50% of vote at the last four elections (left hand scale) and Liberal Democrat support (right hand scale)

25%

60% 20% 50% 15%

40% 30%

10%

20%

Liberal Democrat Support (%)

% of seats won with less than 50% of the vote

70%

5% 10% 0%

0% 1997

2001

2005

% seats won on less than 50%

2010

Lib Dem %

Does MPs being elected on less than 50% of the vote matter? The growth of the third and minor parties since the 1970s has greatly diminished the share of the vote held by the biggest two parties, both nationally and locally. But it is not clear that this increase in competitiveness and complexity has actually undermined the status of MPs as representatives of their constituency. MPs being elected on less than 50% of the vote has not normally been an argument for electoral reform. Under most reformers’ preferred system, STV, MPs would have to pass a quota of less than 50% in multi‐ member constituencies, for example in a five member seat the quota would normally be one sixth of the vote or 16.67%. While it is perfectly designed for AV, the 50% argument is not particularly compelling, especially when one considers that the 50% support or acquiescence may well be on third preferences in a competitive seat – hardly a ringing endorsement. Furthermore, a strong argument can be made that this problem is not really a problem at all. MPs should represent all their constituents, regardless of how they voted, and whether or not they voted at all. In Hull West and Hessle, Alan Johnson MP won 42.5% of the vote on a low turnout of 45.6% ‐ which equates to slightly less than 20% of the electorate. Yet to suggest that he does not represent the other 80% would be quite offensive, especially in a relatively deprived seat with a substantial casework load.

Under this argument, MPs would only heed the opinions of those who supported them or acquiesced to their election. This argument belittles the role of an MP as representative of their entire constituency and all its constituents and should be dismissed. Would AV make a difference? By enabling voters to express second and third preferences, AV does force MPs to achieve 50% of the valid votes left after each round of reallocations. This is not exactly 50% of the vote unless ranking all candidates is compulsory, since votes drop out of the counting if they have no preferences for candidates remaining (e.g. if a voter votes BNP and no‐one else, once the BNP are eliminated the vote is no longer considered valid). While it is described as 50% of the support, it is really 50% of a forced choice between the top two candidates, with ‘don’t knows’ discarded. The Jenkins Commission noted that AV “would virtually ensure that each MP commanded at least majority acquiescence within his constituency”. Note “acquiescence” rather than “support” – a key distinction and one that is often missed.23 If we believe that an MP elected on 51% after third preferences have been reallocated is more legitimate than one elected on 49% actual support, then AV is an improvement. This requires us to treat third preferences equal to first, violating the principle that all voters should be treated equally. Conclusions on MPs being elected on 50% of the vote MPs being elected on less than 50% of the vote is a trivial problem that has only been given prominence since AV was selected for a referendum. It also misrepresents the role of an MP as the representative of his or her constituency, not just the section of the electorate that voted for them. AV’s solution to this non‐problem requires violating equal treatment of voters by treating the acquiescence of a third preference as equivalent to the enthusiastic support of a first preference. AV solves a trivial problem at the cost of undermining democratic legitimacy. Changing to AV could actually make some MPs’ legitimacy more open to question. If an MP were elected only because of third preferences they could be seen as less legitimate than one currently elected on less than 50% of the vote under FPTP.

Summary of FPTP vs. AV Table 3. Scorecard on six issues: Is AV better than FPTP?

Does it matter?

Does AV clearly improve on FPTP?

Disproportionality and perverse results

Depends on beliefs about No democracy; it does for some

Safe seats

Yes

No ‐ there are better ways to tackle seats for life

Decisive Results

Yes

Not necessarily – may increase the bias to Labour producing huge landslides on one side and only coalitions on the other

Wasted Votes

No – votes cast for losing AV still wastes lots of votes. Since recycled votes candidates do count, they are more undemocratic than wasted ones, it’s just don’t win probably worse

Tactical Voting

Depends on your view of It formalises tactical voting so everyone does it. voters’ motives Whether this is an improvement depends on your attitude to tactical voting

MPs elected on less than Only since AV was put on the Yes 50% of the vote table The six issues examined in detail do not show that AV is clearly superior to FPTP. Only in one trivial case – MPs being elected on less than 50% of the vote – is it clearly superior, and the solution leads to MPs being elected on acquiescence as well as support. Disproportionality and safe seats are not clear wins for either system – AV can be extremely disproportional in some circumstances, while safe seats are totally unaffected by it. On decisive results, AV may exaggerate swings which would exacerbate the biases already in the system, producing absurdly large majorities for Labour but making it harder for other parties. On wasted votes and tactical voting, there is an ideological argument to make either way, and this probably depends on prior preferences about electoral systems. But to declare AV superior you do have to make some significant assumptions about voters’ motivations which are not convincing. On these six tests it is clear that AV is not an improvement on FPTP. The case for changing the electoral system from FPTP to AV is extremely weak.

The slippery slope to STV Some supporters of electoral reform are supporting AV on the basis that it will lead to a more proportional system later on ‐ The Electoral Reform Society view AV as “a logical progression” to their preferred system, STV.24 That the strongest argument for AV is that it will easily be replaced should serve to highlight its fundamental lack of appeal. AV is equivalent to STV with single‐member seats, so could be seen as a viable halfway house between FPTP and STV. Figure 4 shows the historic and potential trajectory of reform. The argument runs thus: in the coalition talks after the general election AV was the only system on offer. However, under AV coalitions are more likely and so the leverage of third parties becomes stronger. Since

third and minor parties would benefit most from STV and it would be more acceptable than list‐based systems, their increased power makes its introduction more likely under AV. Figure 4: How we got here, and could end up with STV Simple vo�ng

Preferen�al Vo�ng

Mul�-member cons�tuencies

19th century: FPTP, monstly with 2-member seats

Eventually? STV with mil�-member seats

Single-member cons�tuencies

Now: FPTP with single member seats

Soon? AV with single member seats

It is disturbing that the strongest argument for AV is that it will be easily replaced. The uncertainty around how voters and the parties would react to AV means that we would risk ending up permanently with AV, a system no‐one actually wants as their first preference. It also risks ending up with STV, a system where the constituency link is diluted to the point of being destroyed, MPs can be elected with much less support than is the case now and where extreme parties have a much better chance of being elected. A change to the electoral system is a major reform and should be assessed on the merits of the systems under consideration, not fantasies of further change.

Candidates FPTP allows ideologically diverse candidates to prosper by encouraging the parties to appeal to a broad cross‐ section of opinion. Other systems do not encourage this, especially list PR (used to elect MEPs) where parties’ central organisations have very significant control over who gets selected and therefore elected. FPTP has long been criticised by feminists for being particularly hard for women and people from ethnic minorities to succeed in.25 Both Labour and the Conservatives have made very significant progress on this under FPTP, changing the rules to encourage or force local parties to select more representative candidates (in terms of gender and ethnicity at least).26 While list systems, especially closed lists, make positive discrimination easier, AV does not do this any more than FPTP – in fact, if local parties are seeking the most broadly acceptable candidates they might shy away from selecting candidates who are women or from an ethnic minority in favour of a “most acceptable” candidate. AV’s promotion of the least‐disliked candidate risks leading to a bland politics where politicians are afraid to express any controversial opinions or contrary positions.

Why focus on the Commons? The House of Commons is, of all the legislative bodies that govern us, the least in need of reform. The Closed List PR system for the European Parliament and especially the unelected House of Lords could be argued to be far worse than FPTP in terms of safe seats and constituency links. The Closed List PR system used to elect MEPs since 1999 manages to make MEPs remote from their electorate, and in some cases very hard to dismiss – probably the safest seat in British politics is the number one spot on any of the Conservative, Labour, Liberal Democrat or UKIP party lists in London or the Southeast, the largest regions for this election – in the Southeast in 2009 Labour saw one MEP elected having received only 8.2% of the vote. A life peerage is by definition the ultimate safe seat. The electoral system for the proposed elected House of Lords would seem to be a much better debate to have than whether to have a minor change to the House of Commons or not, since it is a massive change to a clearly undemocratic situation.

Conclusions All electoral systems are imperfect compromises between things we hold important. FPTP disregards proportionality in favour of the constituency link and the ability to “throw the rascals out”; proportional systems make the reverse choice. AV is a slight modification of FPTP which maintains the constituency link, but modifies how we elect MPs. It might make single‐party government less likely, although Australia does not suggest that this outcome is certain. In a referendum choosing between AV and FPTP, the arguments should be quite narrow – the two systems are not very different after all. A change would make no difference to the existence of safe seats or the under‐ representation of women or people who are members of ethnic minorities. AV can be extremely disproportional – more so than FPTP. Tactical voting changes under AV, by formalising and effectively encouraging it. Voters can have their cake and eat it too under AV. The important question in deciding between the two systems are whether we find a losing vote (described as a “wasted vote”, although this is a problematic term) more undemocratic than a “recycled vote” under AV; whether we believe that AV would make it harder to “throw the rascals out” by creating permanent coalitions out of our current three main parties; and whether it is a slippery slope to STV or list PR. Recycled votes are more undemocratic than wasted votes. AV treats a third preference as equal to a first, in order to solve a problem that does not matter – MPs being elected on less than 50% of the vote. That AV was introduced in Australia to allow anti‐Labor voters to vote for several anti‐Labor parties is not an inspiring basis on which to reform our democracy. It is impossible to know how the parties and voters will react to AV if it comes in. Politics is in almost every democracy a contest between two viable options: a left‐of‐centre and right‐of‐centre government, and AV

would not change this fundamentally. But it would not improve the ability of the electorate to remove a government – it might make no difference or diminish it. This is not a compelling reason to change voting system. There is a risk that changing to AV could be a slippery slope towards one of two electoral systems that dilute or destroy the constituency link: STV or list PR. Both of these would be extremely radical departures from single‐ member constituencies with clearly accountable MPs, and should be resisted at all costs.

Endnotes

HC Debs, 5th July 2010, col. 24 All references in this section from the Jenkins Commission, accessed from http://www.archive.official‐ documents.co.uk/document/cm40/4090/contents.htm 3 Jenkins Commission 4 Reilly,B., 'The Alternative Vote and Ethnic Accommodation: New Evidence from Papua New Guinea'. Electoral Studies, Vol. 16(1), 1997 pp1‐11; Fraenkel , J and Grofman, B (2006) “Does the Alternative Vote foster moderation in ethnically divided societies? The case of Fiji” Comparative Political Studies, 39, 5, pp623‐651 5 Bennett, S and Lundie, R “Australian Election Systems” p8 6 Adam Carr’s Election Archives, http://psephos.adam‐carr.net/countries/a/australia/1980/1980repsnsw.txt, quoted in Bennett and Lundie 7 Hix, S, Johnsonton, R and MacLean, I “Choosing an Electoral System”, British Academy, p53. Thanks to Chris Gatenby for pointing out some cases where this was the case in 2010 in Western Australia 8 In 1951, counting the National Liberals together with the Conservatives produces ratios of 0.96 for Labour and 1.07 for the coalition; without counting them together 1951 is also a case of the losing party (Labour) having a higher ratio. 1992 was extremely close: the Conservatives won 1.23 seats for every 1% of the vote, Labour 1.21 seats. 9 Denver, D., et al. (1998). "Triumph of Targeting". British Elections and Parties Review 8:171‐186; Evans, G., et al. (1998). "New Labour, New Tactical Voting?" British Elections and Parties Review 8.Johnston, R., et al. (2001). From Votes to Seats. Manchester: Manchester University Press; Pattie, C., and R. Johnston. (2003b). "Local Battles in a national landslide: constituency campaigning at the 2001 British General Election". Political Geography 22:381‐ 414; Whiteley, P., and P. Seyd. (2003). "How to win a landslide by really trying: the effects of local campaigning on voting in the 1997 British General Election". Electoral Studies 22:301‐324. 10 In the 1998 German elections, Helmut Kohl lost his constituency seat in Ludwigshafen by 7.1% of the vote but returned to the Bundestag via the party list for the Rheinland‐Pfalz region. 11 Dunleavy, P, Margetts, H, O’Duffy, B, Weir, S (1998) ‘Remodelling the 1997 General Election: How Britain would have voted under alternative electoral systems’ Journal of Elections, Public Opinion and Parties, 8, 1, pp208‐231, p225 12 Hix, S, Johnsonton, R and MacLean, I “Choosing an Electoral System”, British Academy, p52 13 Sanders, D, Clarke, H, Stewart, M and Whitely, P (2011, forthcoming) ‘Simulating the Effects of the alternative vote in the 2010 UK general election’, Parliamentary Affairs 14 Although the notional figure for Liverpool Wavertree in 2005 for the 2010 boundaries had a Labour share of 49% 15 http://www.takebackparliament.com/sites/takebackparliament/index.php/pages/about‐this‐campaign 16 Sanders et al (2011) 17 Bolton West, Huddersfield West, Cardiganshire and Carmarthenshire 18 HM Government (2010) ‘The Coalition: Our programme for Government’ p27 19 Goldsmith, Z (2010) “We were promised we'd be able to recall MPs...so much for the Coalition's promises!” Mail on Sunday 8th August 20 Electoral Reform Society http://www.electoral‐reform.org.uk/article.php?id=54 21 Riker, W and Ordeshook (1968) P “A Theory of the Calculus of Voting” American Political Science Review 62 (1), pp25‐42; see also the massive literature that follows this debate which is still unresolved today 22 Electoral Reform Society 23 Jenkins Commission 24 Electoral Reform Society http://www.electoral‐reform.org.uk/article.php?id=48 25 Lovenduski, J (2005) Feminizing Politics, Cambridge, Polity 26 McIlveen, R (2009) Ladies of the Right: an interim analysis of the A‐list, Journal of Elections, Public Opinion and Parties, 19 (2) pp147‐157 2

2 5

research note October 2010

About the Authors Dr Robert McIlveen completed his PhD in Political Science in 2008 at the University of Sheffield. He has About the Authors interests in statistical methods, game theory and rational choice. He joined Policy Exchange’s Environment and Energy Unit before returning to electoral and constitutional issues. This is the ninth report for Policy Exchange Dr Robert McIlveen completed his PhD in Political Science in 2008 at the University of Sheffield. He has he was written, edited or contributed to. interests in statistical methods, game theory and rational choice. He joined Policy Exchange’s Environment and Energy Unit before returning to electoral and constitutional issues. This is the ninth report for Policy Exchange Natalie Evans is Deputy Director at Policy Exchange, responsible for the strategic direction and output of the he was written, edited or contributed to. research team.

Acknowledgements

The author would like to thank those academics who gave advice, opinions and suggestions in the course of Acknowledgements research, in particular Dr Michael Pinto‐Duschinsky. The author would like to thank those academics who gave advice, opinions and suggestions in the course of research, in particular Dr Michael Pinto‐Duschinsky. About Policy Exchange Policy Exchange, an independent educational charity. Our mission is to develop and promote new policy ideas which will foster a free society About Policy Exchange based on strong communities, limited government, national self

confidence and an enterprise culture. In contrast to many other think tanks Policy Exchange is Policy Exchange, an independent educational charity. Our mission is to develop and promote new policy committed to an evidence‐based approach to policy development. Our impact speaks for itself: from ideas which will foster a free society based on strong communities, limited government, national self housing to policing reform, education to the NHS, our proposals have been taken on board by the main confidence and an enterprise culture. In contrast to many other think tanks Policy Exchange is political parties. Registered charity number 1096300. committed to an evidence‐based approach to policy development. Our impact speaks for itself: from housing to policing reform, education to the NHS, our proposals have been taken on board by the main For more information please contact us at: Clutha House, 10 Storey’s Gate, London, SW1P 3AY. political parties. Registered charity number 1096300. Alternatively we can be contacted on 020 7340 2650 and at info@policyexchange.org.uk For more information please contact us at: Clutha House, 10 Storey’s Gate, London, SW1P 3AY. Alternatively we can be contacted on 020 7340 2650 and at info@policyexchange.org.uk

Trustees

Charles Moore (Chairman of the Board), Richard Ehrman (Deputy Chair), Theodore Agnew, Richard

Briance, Camilla Cavendish, Robin Edwards, Trustees Virginia Fraser, George Robinson, Robert Rosenkranz, Andrew Sells, Tim Steel, Alice Thomson, Rachel Whetstone, Simon Wolfson Charles Moore (Chairman of the Board), Richard Ehrman (Deputy Chair), Theodore Agnew, Richard Briance, Camilla Cavendish, Robin Edwards, Virginia Fraser, George Robinson, Robert Rosenkranz, Andrew Sells, Tim Steel, Alice Thomson, Rachel Whetstone, Simon Wolfson